Google · model
Gemini 3 Pro benchmark scores
As of 2026-10-07, Gemini 3 Pro (Google) has measured scores on 22 benchmarks and estimated scores on 74 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 90.8% | measured |
| AA-HLE | 39.7% | measured |
| AA-SciCode | 47.8% | estimated ± 3.8 pp, high confidence |
| AIME26 | 100.0% | estimated ± 0.6 pp, low confidence |
| Artificial Analysis Intelligence Index | 28.0% | measured |
| CritPt | 9.1% | measured |
| GDPval-AA | 21.9% | estimated ± 6.2 pp, medium confidence |
| GPQA Diamond | 86.3% | estimated ± 1.4 pp, medium confidence |
| HLE | 43.9% | estimated ± 6.3 pp, medium confidence |
| IFBench | 69.0% | estimated ± 6.9 pp, medium confidence |
| AA-LCR | 76.0% | measured |
| AA-Omniscience Accuracy | 55.8% | measured |
| BrowseComp | 81.1% | estimated ± 5.7 pp, medium confidence |
| IFEval | 93.4% | estimated ± 1.3 pp, high confidence |
| LongBench v2 | 61.5% | estimated ± 5.9 pp, medium confidence |
| τ²-bench results | 87.1% | measured |
| AA Coding Index | 46.4% | estimated ± 6.8 pp, medium confidence |
| AA-MMMU-Pro | 80.2% | measured |
| APEX-Agents-AA | 15.5% | estimated ± 3.6 pp, medium confidence |
| BioMysteryBench (human-difficult) | 47.9% | estimated ± 5.9 pp, low confidence |
| DeepSearchQA | 73.4% | estimated ± 10.9 pp, low confidence |
| SWE-bench Verified | 74.4% | estimated ± 4.4 pp, high confidence |
| BFCL v4 | 62.6% | estimated ± 9.8 pp, low confidence |
| CyberGym | 65.7% | estimated ± 10.0 pp, low confidence |
| JobBench | 11.4% | measured |
| τ³-bench results | 68.9% | estimated ± 8.9 pp, medium confidence |
| WideResearch | 72.8% | estimated ± 1.2 pp, medium confidence |
| AA Agentic Index | 26.2% | estimated ± 6.4 pp, medium confidence |
| MMLU-Pro | 85.7% | estimated ± 3.5 pp, medium confidence |
| FrontierMath v2 (Tier 4) | 18.8% | measured |
| FrontierMath v2 (Tiers 1-3) | 37.6% | measured |
| GPQA | 88.0% | estimated ± 1.6 pp, high confidence |
| AA-IFBench | 70.4% | measured |
| Gert Labs | 63.2% | measured |
| AA EnterpriseOps-Gym | 33.1% | estimated ± 3.6 pp, medium confidence |
| ApprenticeBench | 5.4% | estimated ± 3.1 pp, medium confidence |
| ARC-AGI-1 | 75.0% | measured |
| ARC-AGI-2 | 31.1% | measured |
| FrontierCode 1.1 Main | 5.7% | estimated ± 4.3 pp, low confidence |
| OfficeQA Pro | 57.8% | estimated ± 7.2 pp, medium confidence |
| OSWorld-Verified | 73.0% | estimated ± 5.9 pp, medium confidence |
| Vals GPQA Diamond | 85.9% | estimated ± 2.2 pp, high confidence |
| Vals MMLU-Pro | 85.4% | estimated ± 1.8 pp, high confidence |
| Vals SWE-bench | 69.9% | estimated ± 6.3 pp, medium confidence |
| AA AutomationBench | 42.7% | estimated ± 5.0 pp, low confidence |
| GDP.pdf | 17.8% | estimated ± 5.2 pp, low confidence |
| AA ITBench | 51.3% | estimated ± 3.6 pp, medium confidence |
| HLE w/o tools | 38.0% | estimated ± 5.7 pp, medium confidence |
| OSWorld 2.0 | 9.8% | estimated ± 4.5 pp, medium confidence |
| PostTrainBench v1.1 | 19.9% | estimated ± 10.5 pp, low confidence |
| Toolathlon-Verified | 73.2% | estimated ± 1.3 pp, low confidence |
| Vibe Code Bench | 14.3% | measured |
| CharXiv | 81.4% | measured |
| CharXiv w/o tools | 83.7% | estimated ± 1.7 pp, low confidence |
| C-Eval | 93.7% | estimated ± 1.0 pp, low confidence |
| Claw-Eval | 50.7% | estimated ± 5.6 pp, low confidence |
| HMMT Nov 2025 | 93.2% | estimated ± 2.6 pp, low confidence |
| MathVision | 86.6% | measured |
| MCP Atlas | 31.8% | estimated ± 8.2 pp, medium confidence |
| MCP-Tasks | 71.4% | estimated ± 8.2 pp, low confidence |
| MMAnswerBench | 86.2% | estimated ± 1.3 pp, low confidence |
| MMLU-Redux | 97.7% | estimated ± 1.1 pp, low confidence |
| MMMU-Pro | 81.0% | measured |
| ScreenSpot Pro | 72.7% | measured |
| SuperGPQA | 87.7% | estimated ± 7.5 pp, low confidence |
| Toolathlon | 50.1% | estimated ± 7.2 pp, medium confidence |
| V* | 88.0% | measured |
| VideoMMMU | 87.6% | measured |
| ERQA | 64.0% | estimated ± 2.4 pp, low confidence |
| HealthBench Hard | 23.7% | estimated ± 7.7 pp, low confidence |
| MedXpertQA (MM) | 76.6% | estimated ± 2.4 pp, high confidence |
| MMLU-Pro (Arcee) | 100.0% | estimated ± 4.0 pp, low confidence |
| React Native Evals | 61.9% | estimated ± 3.0 pp, low confidence |
| ResearchClawBench | 18.0% | estimated ± 2.2 pp, medium confidence |
| SWE-Rebench | 7.6% | estimated ± 6.5 pp, low confidence |
| ARC-AGI-3 | 1.8% | estimated ± 3.4 pp, high confidence |
| FrontierMath (legacy) | 38.2% | estimated ± 0.7 pp, low confidence |
| BioMysteryBench (human-solvable) | 82.7% | estimated ± 0.6 pp, low confidence |
| HealthBench (raw) | 48.8% | estimated ± 6.4 pp, low confidence |
| HealthBench (length-adjusted) | 48.9% | estimated ± 2.5 pp, low confidence |
| HealthBench Professional | 55.1% | estimated ± 4.6 pp, medium confidence |
| HealthBench Professional (raw) | 59.2% | estimated ± 5.0 pp, medium confidence |
| HLE-Verified | 35.3% | estimated ± 0.7 pp, low confidence |
| LABBench2 | 41.6% | estimated ± 1.3 pp, low confidence |
| MMMU | 87.4% | estimated ± 0.9 pp, medium confidence |
| Agents' Last Exam | 48.9% | estimated ± 8.7 pp, low confidence |
| BabyVision | 48.3% | estimated ± 4.3 pp, low confidence |
| MMVU | 75.9% | estimated ± 2.9 pp, medium confidence |
| SimpleVQA | 85.6% | estimated ± 5.7 pp, low confidence |
| ZeroBench | 13.8% | estimated ± 4.7 pp, low confidence |
| MMLU | 95.6% | estimated ± 1.1 pp, low confidence |
| MMMU-Pro w/ Python | 82.4% | estimated ± 0.6 pp, high confidence |
| MathVision w/ Python | 93.2% | estimated ± 0.5 pp, low confidence |
| RealWorldQA | 3.9% | estimated ± 0.8 pp, low confidence |
| Video-MME (with subtitle) | 87.7% | estimated ± 0.9 pp, medium confidence |
| OmniDocBench 1.5 | 91.7% | estimated ± 5.9 pp, low confidence |